Download Analysis of Sound in the Mammalian Ear: A History of Discoveries

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Sound wikipedia , lookup

Sound localization wikipedia , lookup

Sound from ultrasound wikipedia , lookup

Sensorineural hearing loss wikipedia , lookup

Auditory system wikipedia , lookup

Transcript
Syracuse Scholar (1979-1991)
Volume 3
Issue 1 Syracuse Scholar Spring 1982
Article 9
1982
Analysis of Sound in the Mammalian Ear: A
History of Discoveries
]ozef J. Zwislocki
Follow this and additional works at: http://surface.syr.edu/suscholar
Recommended Citation
Zwislocki, ]ozef J. (1982) "Analysis of Sound in the Mammalian Ear: A History of Discoveries," Syracuse Scholar (1979-1991): Vol. 3:
Iss. 1, Article 9.
Available at: http://surface.syr.edu/suscholar/vol3/iss1/9
This Article is brought to you for free and open access by SURFACE. It has been accepted for inclusion in Syracuse Scholar (1979-1991) by an
authorized administrator of SURFACE. For more information, please contact [email protected].
Zwislocki: Analysis of Sound in the Mammalian Ear: A History of Discoveries
37
Analysis of Sound
in the
Mammalian Ear:
A History of Discoveries
]ozef]. Zwislocki
I
Jozef]. Zwislocki is Professor of Sensory
Sciences and Director of the Institute
for Sensory Research at the L.C. Smith
College of Engineering, Syracuse U niversity. After receiving his doctorate
from the Federal Technical Institute in
Zurich, he served as head of the Electroacoustic Laboratory at the University of
Basle and spent six years as a Research
Fellow at Harvard's Psychoacoustic
Laboratory. Dr. Zwislocki's current
interests focus on the biophysics of the
cochlea and measurement of subjective
sensory impressions.
Dr. Zwislocki's article is a slightly
modified version of an article that
originally appeared in American
Scientist 69 (1981): 184-192. We are
grateful to the publishers of American
Scientist for permission to adapt their
material.
Published by SURFACE, 1982
n vision, when two wavelengths of visible light are combined, the
sensation of a single color arises: A mixture of blue and yellow
makes a green that is indistinguishable from the green resulting
from one appropriate wavelength. In hearing, a mixture of two sound
waves of different lengths never produces the same impression as one
wave; when the difference between the wavelengths is sufficient, two
distinct tones are heard. Rather than present to our consciousness the
impression of a single sound, the ear decomposes complex sounds into
their components. In fact, its analytic power is so great that we are able
to pick single instruments out of a whole orchestra and to detect
changes in tone frequency as small as 0.3%.
How is the ear capable of this feat? From centuries of arduous research we have gained some idea, but the answer is not yet complete.
Current laboratory work is still in hot pursuit of certain crucial aspects
of that answer. To the surprise of many, it appears that the main auditory analysis of sounds is performed almost entirely by mechanical
rather than by electrical or neural means.
Our present knowledge of auditory sound analysis has emerged
gradually, over centuries, but at an accelerating pace, from several
scientific disciplines including physics, anatomy, physiology, and
psychology. About 530 B.C. Pythagoras observed that musical pitch is
uniquely related to the length of a vibrating string, and that halving the
length produces an octave interval. History waited over a thousand
years for a further insight by Galileo: The length of a string determines
its natural vibration frequency, and the sensation of pitch depends on
the frequency rather than directly on the string's length. The two
parameters of length and frequency can be separated, since the natural
frequency depends on a combination of length, mass, and tension.
Shortly after Galileo's insight, the phenomenon of resonance was
discovered: A sound corresponding to the natural frequency of a string
was seen to cause the string to vibrate; other frequencies did not have
this capacity. A string, it was further found, has several natural
1
Syracuse Scholar (1979-1991), Vol. 3, Iss. 1 [1982], Art. 9
frequencies which are multiples of each other; these have come to be
known as harmonics. When a string is made to vibrate with several
harmonics simultaneously, its pattern of vibration depends on their
number and relative amplitudes.
About 150 years later, Fourier discovered that any desired pattern of
vibration can be obtained by an appropriate composition of harmonics;
conversely, any single-valued function can be decomposed into a set of
harmonics, described mathematically as sinusoids. In the same period,
Ohm, whose name became attached to a well-known law of electrical
circuits, observed that the ear performs a Fourier analysis of complex
sounds-in other words, that we are able to hear their sinusoidal
components.
It seemed plausible to conjecture that this analysis of sounds resulted
from resonance within the ear. The ear would then have to contain a set
of resonators, each tuned to a different frequency. The search for the
resonators was on. However, the search required a level of knowledge
of the ear's anatomy that was not reached until the middle of the nineteenth century.
Figure 2. Schematic cross section of a
single turn of the cochlea. Oval dotted
lines indicate the area seen in more
detail in Figure 3. Points x andy here
correspond to x andy in Figure 3.
The search for resonators
It has been known as early as the seventeenth century that sound
entering the auditory canal and striking the tympanic membrane was
led through a set of small bones in the middle ear to a labyrinth of bony
canals constituting the inner ear (see Fig. 1). One of these canals,
outer ear
inner ear
Figure 1. Sound first passes through the outer ear, striking the tympanic membrane.
The resulting vibration then enters the middle ear, a chamber containing three small
ossicles, or bones. These ossicles (malleus, incus, and stapes) transmit sound to the inner
ear, which consists of the cochlea and the semicircular canal system. Sound analysis occurs
in the cochlea.
http://surface.syr.edu/suscholar/vol3/iss1/9
2
ANALYSIS OF SOUND- 39
Zwislocki: Analysis of Sound in the Mammalian Ear: A History of Discoveries
~-------------------------------------------------------------.y
X~---------------------------------------------------------------------1
Figure 3. Schematic cross section of a single turn of the cochlea shows details of its struc·
ture. Note the relative positions of the three main canals within the cochlea: the scala
vestibuli, scala tympani, and cochlear duct. Compare with Figure 2.
shaped like the spiral shell of a snail and called the cochlea (from the
Greek word for snail), was regarded as the organ of hearing. The
cochlear canal is divided longitudinally by a narrow partition that
consists of two strips, one bony and one membranous. Only the bony
part, the spiral lamina, was known in the seventeenth century. It is
realtively wide near the cochlear base-the point where sound enters
the cochlea-and becomes progressively narrower toward the cochlear
apex.
Duverney, a famous French anatomist of the seventeenth century,
suggested that the spiral lamina might act as a set of resonators,
responding to low frequencies at the cochlear base and to high frequencies at the cochlear apex. However, improved anatomical techniques of the nineteenth century revealed that the membranous part of
the cochlear partition, the cochlear duct, is the essential component,
since it supports the sensory cells connected to the fibers of the auditory
nerve. The duct is narrow at the cochlear base and wide at the apex.
Accordingly, Duverney's idea was turned around; it was hypothesized
that the resonators for high frequencies were located at the cochlear
base, those for low frequencies at the apex.
Two schematic drawings of the cochlea based on modern anatomicai
knowledge show its structure in more detail (Figs. 3 and 4). The
Published by SURFACE, 1982
3
Syracuse Scholar (1979-1991), Vol. 3, Iss. 1 [1982], Art. 9
cochlear duct divides the cochlea longitudinally into two parts-an
upper portion, called the scala vestibuli, and a lower portion, the scala
tympani, both filled with perilyrpph related to the cerebrospinal fluid.
The cochlear duct, or scala media, itself is filled with endolymph, a fluid
similar in ionic composition to intracellular fluid. Sound enters the
cochlea through an opening in the bony wall of the basal part of the
scala vestibuli known as the oval window and leaves it through the
round window, an opening in the basal wall of the scala tympani. A
third opening, the helicotrema, equalizes static pressure between the
scala vestibuli and the scala tympani. The fundamental configuration is
clear from the simplified model of an uncoiled cochlea shown in Figure 3.
The scala media is separated from the scala vestibuli by a thin vestibular membrane known as Reissner's membrane, and from the scala
tympani by the complex and much thicker basilar membrane and the
spiral lamina. The organ of Corti, a cell mass resting on the basilar
membrane, contains the stiff arches of Corti as well as the auditory
sense cells, called hair cells because of the numerous cilia protruding
from their apices. The hair cells are arranged in four rows running the
length of the cochlea-three rows of outer hair cells on the outside of
the arches of Corti, and a fourth row, containing the inner hair cells, on
the inside (see Fig. 4). The two types of hair cells are quite different
from each other and seem to perform different functions. All the hair
cells are innervated by the afferent neurons of the auditory nerve,
whose terminal fibers can be seen as a long striation through the spiral
lamina. Interestingly, 90 % to 95 % of the fibers end on the less
numerous inner hair cells; only 5% to 10% terminate on the outer hair
cells, each innervating several of them.
spiral limb\
Helmholtz's theory
The current concept of how the cochlea analyzes sound has its roots
in the resonance theory proposed by Helmholtz in the middle of the
nineteenth century. His work coincided with rapid progress in cochlear
anatomy, and he modified his theory as new knowledge became available. At first Helmholtz thought that the arches of Corti acted as
Figure 4. Longitudinal section of uncoiled cochlea (simplified version) shows the two
''windows" (oval and round) by which sound enters and leaves the organ. Note the wave
pattern drawn according to Bekesy's observations.
http://surface.syr.edu/suscholar/vol3/iss1/9
4
ANALYSIS OF SOUND-4 1
Zwislocki: Analysis of Sound in the Mammalian Ear: A History of Discoveries
0
.0
"'
:J
2.
0
0
a:
Figure 5. Cross section through the organ of Corti and the rectorial membrane near the
cochlear apex illustrates the anatomical rel ationship of these two bodies, nearly equal in
area. Note the three rows of outer hair cells, attached to the rectorial membrane by their
ste reocilia, and one row of inner hair cells.
resonators. Because the size of the arches increases with the distance
from the cochlear base, he concluded that the base of the cochlea
responded to high frequencies, the apex to low frequencies-just the
opposite of what Duverney had assumed. When it became known that
the basilar membrane is narrow at the base, gradually widening toward
the apex, and that it contains fibers that span its width like the strings
of a musical instrument, he abandoned the arches of Corti as possible
resonators in favor of the fibers.
According to Helmholtz's theory, sound is transmitted to the basilar
membrane through the perilymph of the scala vestibuli and the endolymph of the cochlear duct, producing maximum vibration at the place
where the basilar membrane is tuned to the particular frequency of the .
sound. When the sound contains several frequencies , each produces a
separate vibration maximum at a different location of the basilar
membrane. In this way the frequency spectrum is resolved spatially
along the cochlea. Extending Johannes Muller's doctrine of specific
nerve energies, Helmholtz proposed that the cochlear nerve fibers
ending at the location of maximum vibration were maximally excited
and determined the pitch sensation. Thus excitation of the fibers in the
Published by SURFACE, 1982
5
Syracuse Scholar (1979-1991), Vol. 3, Iss. 1 [1982], Art. 9
cochlear base by high frequencies produced a sensation of high pitch,
and excitation of the fibers near the apex by low frequencies gave rise to
a sensation of low pitch. Because of this relation between pitch and
cochlear location, Helmholtz's theory is now often called the place
theory.
Helmholtz's resonance theory did not answer all the relevant questions. It stirred up a prolonged controversy leading to numerous
counterproposals and experiments with mechanical models that more
or less reproduced the mechanical features of the cochlea. Finally,
Georg von Bekesy, a Hungarian physicist, laid most of the controversy
to rest in the first half of this century with the help of ingenious experiments making use of cochlear models and postmortem preparations.
His results indicated that the place theory was correct but that the local
vibration maximum of the basilar membrane was not due to ordinary
resonance, as Helmholtz had imagined.
Bekesy's experiments
In a series of experiments Bekesy constructed enlarged models of the
human cochlea, attempting to preserve the proportions among the
various cochlear dimensions as closely as possible. To simplify both the
construction of the model and the observation of vibration, he did not
attempt to duplicate the cochlear spiral but straightened it out, obtaining the configuration shown in Figure 4.
Bekesy replaced the cochlear duct with a rubber membrane. The
model was then filled with water, whose viscosity could be varied by
means of glycerin and whose motion was made visible by the addition
of carbon dust. Like the basilar membrane, the rubber membrane increased in width between the basal end of the model and its apex. Sound
entered the model through vibration of a miniature piston suspended in
a small opening that simulated the cochlear oval window. The resulting volume displacements of water were equalized through another
opening placed on the opposite side of the partition, the equivalent of
the round window.
Bekesy observed that sound produced transversal waves on the partition, and ihat the location at which _these waves reached an amplitude
maximum depended on the frequency of the sound. At high frequencies
the amplitude maximum occurred near the windows, gradually moving
toward the other end of the model as the frequency was decreased-just
as Helmholtz had predicted. However, in their movement toward the
apex, the waves traveled right through the amplitude maximum
without any apparent reflection-a pattern that seemed to Bekesy to be
incompatible with Helmholtz's res.o nance theory. The mechanism that
accounted for the amplitude maximum appeared unclear to him and
remained so for many years, although several theoreticians attempted
to explain it.
Empirical confirmation of Bekesy's model experiments did not
become possible until almost twenty years later. During the intervening period he succeeded in improving dissection and microscopic
techniques to the point where he was able to observe the waves in postmortem preparations of human cochleas-work that eventually earned
him the 1961 Nobel Prize. Using amplitude and phase measurements,
he reconstructed the cochlear wave pattern, an example of which
http://surface.syr.edu/suscholar/vol3/iss1/9
6
ANALYSIS OF SOUND-43
Zwislocki: Analysis of Sound in the Mammalian Ear: A History of Discoveries
appears in Figure 4. The systematic measurements had to be limited to
apex of the cochlea, however, where the low frequencies have their
amplitude maxima. Another twenty years passed before the fundamental features of Bekesy' s findings could be confirmed for the
cochlear base.
Before his direct observation of the wave pattern, Bekesy measured
some essential physical parameters of the cochlea, especially the compliances of the membranous parts of the cochlear duct. Finding the
stiffness of the basilar membrane to be the greatest, and that of
Reissner's membrane negligible, he suggested that the basilar membrane effectively determined the elastic properties of the duct, which he
often called (after Helmholtz) the cochlear partition. According to his
measurements, the compliance of the basilar membrane increased from
the base to the apex by between two and three orders of magnitude,
apparently as a result of the membrane's increasing width and
decreasing thickness.
Fundamental mathematical analysis
Bekesy's measurements of the cochlear constants, together with the
known cochlear anatomy, made it possible for me to derive the cochlear
waves mathematically and so to explain their nature. According to
classical hydrodynamic theory, cochlear waves belong to the class of
surface waves that can be observed on the free surface of water. In the
case of the cochlea, the waves arise through the interaction between the
stiffness of the basilar membrane and the inertia of the cochlear fluids;
this interaction provides the longitudinal coupling necessary for wave
propagation.
The elastic longitudinal coupling within the basilar membrane was
found to be negligible because of the narrowness of the basilar
membrane relative to the wavelength. It should be mentioned at this
point that the name basilar membrane is quite misleading. The membrane behaves mechanically like a plate, as Bekesy has demonstrated
and as its morphological structure suggests; but the name was given
before these things became known, and biologists tend to call thin
partitions membranes.
According to my theory, the increasing compliance of the basilar
membrane from the cochlear base to the apex should cause the length of
the cochlear waves to decrease and their amplitude to increase with the
distance from the base, in agreement with Bekesy's observations. If the
membrane and the cochlear fluids were inviscid and no friction intervened, the amplitude would increase all the way to the cochlear apex,
where a wave reflection would take place, leading to standing waves.
However, Bekesy had found a strong damping of transient oscillations of the basilar membrane, indicating substantial friction. The
theory says that friction should gradually convert the mechanical
energy of the waves to heat, as the waves are propagated toward the
cochlear apex. The process is accelerated as the compliance increases,
and the effect of friction gains in relative importance. Ultimately, the
energy loss causes the wave amplitude to decrease. The location at
which this begins to occur depends on sound frequency, since the effect
of stiffness is inversely proportional to the frequency. As a consequence,
the higher the frequency, the closer the point of vibration maximum to
Published by SURFACE, 1982
7
44---SYRACUSE 50-IOLAR
Syracuse Scholar (1979-1991), Vol. 3, Iss. 1 [1982], Art. 9
the cochlear base. The theory shows that, in the presence of strong
friction, the vibration maximum has to occur before the waves reach the
location of basilar-membrane resonance for a given frequency. At the
resonance location the wave amplitude is already so small that it cannot
be measured.
The theoretical picture sketched above provided a sufficient explanation of the wave pattern seen by Bekesy and also accounted for the
fact that he could not detect any resonance phenomena. Most importantly, it defined the relation between sound frequency and the location
of maximum vibration in the cochlea. This relationship was confirmed
by a series of measurements.
Bekesy's measurements, together with my theory, appeared to
account satisfactorily for the fundamental mechanical process responsible both for sound analysis in the cochlea and for the pitch scale. In the
1950s, measurements of cochlear potentials with pairs of electrodes,
one placed in the scala vestibuli and one in the scala tympani, seemed to
fit the same picture, in that the distribution of the potentials along the
cochlea approximated the distribution of mechanical amplitudes found
by Bekesy. It was believed that these potentials were generated by the
hair cells and that they represented the state of excitation of these cells.
The picture appeared satisfactorily cohesive, even though recordings
from single nerve units of the auditory system, which became possible
for the first time in the 1940s, indicated a much sharper frequency
analysis. This was not disturbing, however, since Bekesy had advocated
the possibility of a neural refinement of the coarse mechanical sound
analysis in the cochlea as early as the 1920s, envisioning a mechanism
similar to that which produces contrast effects and the well-known
Mach bands in vision.
New insights into cochlear mechanics
By the late 1960s, however, the picture began to show deep cracks.
Applying the Mossbauer technique, based on absorption of gamma rays,
Johnstone and Boyle succeeded in measuring the vibration of the basilar
membrane in live guinea pigs. They discovered that the amplitude
maximum was considerably sharper than what Bekesy had seen in postmortem preparations. At first their discovery was received with
disbelief, both because of Bekesy's stature and because such a sharp
maximum seemed incompatible with biological tissues, which usually
exhibit high damping. Nevertheless, the findings were soon confirmed
by Rhode on squirrel monkeys. Rhode also compared the amplitude
maximum in live monkeys with that obtained postmortem and found
striking discrepancies .
Rhode discovered other interesting features of cochlear mechanicsin particular, a small notch in the amplitude pattern above the
frequency of the vibration maximum, as well as an infinite wavelength
beginning in the same frequency region. The infinite wavelength indicates, according to mathematical theory, that the resonance frequency
of the basilar membrane has been exceeded. Rhode also found that the
amplitude maximum became flatter as the sound input to the cochlea
increased. This suggests that the damping of the basilar membrane
increases with amplitude.
http://surface.syr.edu/suscholar/vol3/iss1/9
8
ANALYSIS OF SOUND-45
Zwislocki: Analysis of Sound in the Mammalian Ear: A History of Discoveries
Since Rhode could not measure the vibration of the basilar membrane at very low amplitudes compatible with the threshold of
audibility and measurements of neural sensitivity, speculation arose
that the basilar membrane may be as sharply tuned as the auditory
neurons at very low vibration amplitudes. If this were true, the neural
sharpening suggested by Bekesy would be unnecessary. The speculation
was strengthened by direct recording of receptor potentials of the inner
hair cells pioneered by Russell and Sellick in 1977. They not only have
shown that the hair cells are as sharply tuned as the neurons but also
have concluded, on the basis of resistance measurements of the cells,
that the sharp tuning is produced by mechanical means. This discovery
eliminated several hypotheses of neural or electrophysiological sharpening of cochlear frequency analysis .
The question remains, nevertheless, whether the basilar membrane
is indeed as sharply tuned as the hair cells, or whether some mechanical
process taking place between the membrane and the hair cells provides
the additional sharpening. A series of experiments performed on
alligator lizards by scientists at the Massachusetts Institute of Technology suggests that such additional sharpening may take place. The
lizard's basilar membrane does not provide any frequency analysis, yet
its hair cells and the neurons that innervate them do, achieving a tuning
only moderately less sharp than that found in mammals.
The search continues
Suspicion that some mechanical process m the cochlea had been
overlooked also arose from mathematical considerations. Even the
most refined theories of cochlear mechanics were not able to match the
wave patterns found empirically by Rhode. The solution to the
dilemma may be hiding in the structure of the organ of Corti and its
micromechanics (Fig. 5). The tectorial membrane-which is not a
membrane at all but rather a viscoelastic plate of varying thickness
consisting mainly of protein protofibrils-is suspended over the cell
mass of the organ of Corti. It is attached to the outer hair cells through
their hairs, or stereocilia. (Whether the tectorial membrane is also
attached in this fashion to the inner hair cells is not yet clear.)
Since the beginning of this century it has been assumed that the hair
cells are stimulated through a shearing motion that occurs between the
tectorial membrane and the upper surface of the organ of Corti, a
motion that must lead directly-or indirectly, through entrainment of
adjacent fluid-to a bending of the stereocilia. The deformation of the
hair cells themselves during this movement is probably small, since
they are held rigidly in the reticular lamina, a stiff, netlike structure
covering the top surface of the organ of Corti.
It has been shown repeatedly that the bending of stereocilia in an
appropriate plane produces excitation of similar hair cells in organs
related to the cochlea. Anatomically, the appropriate plane for the
cochlea is the cross-sectional plane shown in Figure 5. Ter Kuile in 1900
suggested that the motions of the part of the basilar membrane supporting the organ of Corti and of the tectorial membrane could be
viewed in cross section as rotations of stiff beams around their respective axes-the edge of the spiral lamina for the basilar membrane and
Published by SURFACE, 1982
9
46--SYRACUSE ~OLAR
Syracuse Scholar (1979-1991), Vol. 3, Iss. 1 [1982], Art. 9
the edge of the spiral limbus for the tectorial membrane. If the two
membranes rotated around parallel but different axes, a shearing
motion would have to result between them, as is easy to demonstrate by
a simple geometrical construction.
Whereas Ter Kuile's suggestion still appears acceptable for the
basilar membrane, various studies of the tectorial membrane are more
consistent with the assumption that it bends over its whole free section
between the spiral limbus and the inner hair cells. If this is true, the
shearing motion of the tectorial membrane relative to the organ of
Corti must depend on its stiffness and on the stiffness of its attachments to the upper surface of the organ. It is possible to show by reasonably simple mathematics that the shearing motion almost disappears
when the former stiffness is much smaller than the latter, when the
length of the cochlear waves is sufficiently large, and when the mass of
the tectorial membrane can be disregarded. Such conditions are likely to
prevail at low frequencies ; but at sufficiently high frequencies, the mass
of the tectorial membrane, which is of the same order of magnitude as
the entire cell mass of the organ of Corti, must make itself felt.
If the damping of a mechanical system consisting of a mass attached
to a vibrating support by a stiffness (spring) is small, the system will
resonate when excited at its natural frequency. At resonance, the amplitude of the mass can far exceed that of the support. Since the amplitude
decreases sharply as the frequency changes either upward or downward
from the resonance frequency, a sharp filter effect, or tuning, results . If
the tectorial membrane with its elastic attachments to the organ of
Corti can resonate, a sharply tuned shearing motion between it and the
reticular lamina must result, leading to a sharply tuned excitation of the
inner hair cells.
The idea of such a resonance is supported by a great deal of indirect
evidence. For instance, the stereocilia of hair cells have been found to be
quite stiff and to consist of elastic fibers of actin, the same material that
constitutes muscle fibers. The length of the stereocilia of the outer hair
cells increases from 1 p,m in the cochlear base to 5 p,m or 6 p,m in the
apex. If the mass of the tectorial membrane remained constant, this
factor alone would account for a change in resonance frequency by a
factor of about 25, since the resonance frequency of an elastic beam is
inversely proportional to the square of its length. But the mass of the
tectorial membrane increases considerably toward the apex, further
lowering the resonance frequency. These two factors together could
easily account for the frequency range of over 50 to 1 in the mammalian
cochlea.
Some possible effects of a resonance of the tectorial membrane have
been demonstrated through construction of a simple mechanical model :
A steel reed was clamped rigidly onto the vibrating tip of an electrodynamic vibrator at a small angle to the direction of vibration. The
vibrating tip represented a short section of the basilar membrane; the
reed, a stereocilium. A small brass weight was placed on the tip of the
reed to simulate the mass of the tectorial membrane. Of course, there
are a great many stereocilia in a short section of the organ of Corti; but
this does not affect the essential mechanical relations we are considering, except for increasing the stiffness of the attachment of the
tectorial membrane.
When the vibrator without the reed was excited by an electric current
http://surface.syr.edu/suscholar/vol3/iss1/9
10
ANALYSIS OF SOUND-47
Zwislocki: Analysis of Sound in the Mammalian Ear: A History of Discoveries
of slowly increasing frequency, its amplitude remained constant up to a
cutoff frequency and gradually decreased above that frequency. No
amplitude maximum occurred, because the damping was too high.
When the reed was secured in place, it moved up and down with the
vibrator tip without any amplitude or phase difference until the
frequency approached its transversal-mode resonance. It then began a
strong transversal vibration that reached its maximum at the resonance
frequency. Since the reed's amplitude decreased rapidly on both sides of
the resonance frequency, a sharp tuning resulted. In addition to the
sharp tuning, another, not unexpected, phenomenon was observed.
Near its resonance frequency, the reed appreciably affected the
vibration amplitude of the vibrator tip. Just below the resonance
frequency, this amplitude increased sharply, went through a maximum,
and decreased to a minimum at the resonance frequency. Above the
resonance frequency it showed a small secondary peak before beginning
to decrease in nearly the same manner as it had without the reed. The
amplitude perturbation near the reed's resonance is due to the strong
force the reed exerts on the vibrator during its maximum oscillation.
Depending on the timing of this force, the vibrator amplitude is
increased or decreased.
If the tectorial membrane undergoes a resonance, a similar amplitude
perturbation may be expected in the vibration of the basilar membrane.
Rhode seems to have observed this .
The vibrator model revealed still another effect that has been
repeatedly noted in the cochlea. Near the resonance of the reed a strong
nonlinearity showed up in the wave form of the vibrator. The reason
for this is obvious. It must occur in every coupled system where one part
vibrates at right angles to the vibration of another part. Whether the
reed moves to the right or to the left in the model, it always exerts a
downward force on the vibrator tip. If the reed stood perfectly vertically
on the vibrator tip, the force would be symmetrical and would follow
the second harmonic of the vibrator motion, producing a purely
quadratic distortion. Because the reed stands at a small angle from the
vertical, higher harmonics, especially the third harmonic, arise in
addition to the second.
A similar situation must prevail in the ear if the tectorial membrane
can vibrate nearly orthogonally to the vibration of the basilar membrane, as appears likely. In any event, wave distortions of the type seen
in the vibrator model have been observed in the vibration of the basilar
membrane, in the output of auditory nerve fibers, and in the subjective
responses of listeners.
Although calculations and experiments with models strongly suggest
that resonance of the tectorial membrane system, as outlined above, is
within the realm of possibility, and although a large number of phenomena observed by physical as well as physiological and psychological
methods are consistent with such a resonance, only a direct experiment
can demonstrate this beyond reasonable doubt. Too many parameters
still remain unknown to make any theoretical calculations, or modeling,
conclusive. Neither the internal viscosity of the tectorial membrane,
nor its stiffness, nor the stiffness of its attachments to the organ of
Corti is precisely known. As a consequence, much remains to be
explored, and we are expecting some exciting years of research on
cochlear mechanics.
Published by SURFACE, 1982
11